Re: What about including virus filter to squid ?

On Sat, 2002-11-09 at 04:56, Henrik Nordstrom wrote:
> True.
>
> I don't think inline virus scanning within the Squid process is a good
> path. Instead the virus scanner should be a separate process, possibly
> even on separate machine(s).

It would be 'nice' to allow internal scanning, if desired. But that
should be a matter of appropriate hooks, no more.

> The problem is how a external virus scanner is to integrate with Squid.
> There needs to be something more than just HTTP for this to work in a
> somewhat efficient manner, and it needs to be reasonably standard to
> allow for different virus scanners to be used.
>
> IIRC this is one of the areas where the ICAP protocol can be used. There
> is a ICAP client implementation to Squid (see devel.squid-cache.org),
> and this might allow Squid to integrate with ICAP capable virus scanners
> (there seems to be quite a few already).

Yep. Virus scanning is one of the targets of iCAP.

> Perhaps the ICAP client to Squid hasn't seen as much attention as it
> should?

Definately.

Rob